http://permalink.gmane.org/gmane.comp.python.ipython.user hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://permalink.gmane.org/gmane.comp.python.ipython.user/4187 Hi Benjamin, There definitely is: you can put those lines in your ~/.inputrc and readline will pick them up inside ipython. As a bonus, all other programs that internally use readline will now use vi-mode as well. cheers, Paul Ivanov Benjamin J. Racine, on 2008-12-02 15:45, wrote: Paul Ivanov 2008-12-03T18:18:18 http://permalink.gmane.org/gmane.comp.python.ipython.user/4186 _______________________________________________ IPython-user mailing list IPython-user< at >scipy.org http://lists.ipython.scipy.org/mailman/listinfo/ipython-user Benjamin J. Racine 2008-12-02T23:45:38 http://permalink.gmane.org/gmane.comp.python.ipython.user/4185 Could be, but I really have no idea. On Tue, Dec 2, 2008 at 11:31 AM, mark starnes <m.starnes05< at >imperial.ac.uk> wrote: Brian Granger 2008-12-02T20:39:05 http://permalink.gmane.org/gmane.comp.python.ipython.user/4184 Hi, Could the problem be to do with my mixing of 32 and 64 bit operating systems? I just ran into a problem with a routine that was compiled on the 32 bit machine; it threw an exception on the 64 bit machine.... BR, Mark. Brian Granger wrote: mark starnes 2008-12-02T19:31:34 http://permalink.gmane.org/gmane.comp.python.ipython.user/4183 Hi Brian, On Monday 01 December 2008 19:00:25 Brian Granger wrote: Many thanks for your mail. My code has a class that does a number of things. One of the class methods needs to be run in parallel (à la map()), but I have a problem: <CrunchData is my instantiated class) output = mec.map ( CrunchData.ProcessFile, CrunchData.files) [...] AssertionError: func must be a fuction or str CrunchData.ProcessFile is defined as ( self, fname), so I'm doing something wrong here. If I try execute, and a single file, no luck... pr=mec.execute ('CrunchData.ProcessFile("%s")'%CrunchData.files[0]) pr.r [...] CompositeError: one or more exceptions from call to method: execute [0:execute]: NameError: name 'CrunchData' is not defined [1:execute]: NameError: name 'CrunchData' is not defined [2:execute]: NameError: name 'CrunchData' is not defined [3:execute]: NameError: name 'CrunchData' is not defined [4:execute]: NameError: name 'CrunchData' is not defined [5:execute]: NameError: name 'CrunchData' is not defined [6:e Jose Gómez-Dans 2008-12-02T18:28:24 http://permalink.gmane.org/gmane.comp.python.ipython.user/4182 Hi everyone. I removed Pypar; no effect. The USB control isn't in use AFAIK. I'm not using PBS (though, I hoped to at a later point). I am accessing this program on a remote machine, using $ ssh -X machinename with one ssh for the ipcluster command, and one for the ipython console. The remote machine is suse10.3, 64 bit and is running a vncserver (if this is of any relevance). -------------------------------------- Anyway, I'm finding now, that if I run from IPython command line, I can't get the error to occur. On closing IPython, however, I get: In [4< at >16:28:54]: Do you really want to exit ([y]/n)? y Closing threads... Done. Exception exceptions.TypeError: "'NoneType' object is not callable" in <bound method RemoteReferenceTracker._refLost of <RemoteReferenceTracker(clid=1,url=pbu://127.0.0.1:21778/uwtv36uev6e7emd45xfz77tt75krrduj)>> ignored implying, perhaps that it is occurring but not being reported. -------------------------------------- If I run from the prompt, via $ fe2.py fefile.pyFE I g mark starnes 2008-12-02T15:23:49 http://permalink.gmane.org/gmane.comp.python.ipython.user/4181 I'm not sending signals as far as I know. I will remove all non essential imports in my code, today, to see if it helps. There are two modules I'm not sure of. pypar (my original choice for a python/MPI interface) and a library for controlling hardware via usb. I'll remove them today and see if anything happens. Best regards, Mark. Brian Granger wrote: mark starnes 2008-12-02T09:58:34 http://permalink.gmane.org/gmane.comp.python.ipython.user/4180 Done. https://bugs.launchpad.net/ipython/+bug/304263 for reference. On Dec 1, 2008, at 2:14 PM, Brian Granger wrote: unussum< at >gmail.com 2008-12-02T04:11:15 http://permalink.gmane.org/gmane.comp.python.ipython.user/4179 OK, done! 2008/12/1 Brian Granger <ellisonbg.net< at >gmail.com>: _______________________________________________ IPython-user mailing list IPython-user< at >scipy.org http://lists.ipython.scipy.org/mailman/listinfo/ipython-user Elias Bröms 2008-12-02T03:48:05 http://permalink.gmane.org/gmane.comp.python.ipython.user/4177 Just checked in Twisted and IPython and there are no os.kill's anywhere in the core that would show up here. That leaves: 1. User code 2. Some other process. Is there a chance that the engines are being started using some batch system (like PBS) that ends up sending the process signals? Brian On Mon, Dec 1, 2008 at 2:21 PM, Andrew Straw <strawman< at >astraw.com> wrote: Brian Granger 2008-12-01T23:01:51 http://permalink.gmane.org/gmane.comp.python.ipython.user/4176 Sorry, I wasn't clear here. EINTR is the errno value that's returned here. The signal could be anything. I've never experienced the kernel sending unrequested signals other than the usual things like SIGINT and SIGKILL... But I suppose it could be configured to do so somehow. My best guess is that this is more likely to be in user code than in Twisted or the other sources you list. I suspect that if signals were used as part of Python or Twisted, folks attempting to use non-EINTR-safe code (there's a lot of it) would've come screaming by now. My relatively naive understanding of this stuff is that signals are the blunt tool of inter-process communication and I see the Twisted crew as more the fine surgeon types... But I haven't grepped for os.kill() in the Twisted sources, either, so I could be wrong. Likewise I don't know about IPython, but I'd be surprised if it was doing IPC via signals. If it is, I'd suggest it be removed, as it will come back to haunt anyone without non-EINTR-safe code. It's debuggin Andrew Straw 2008-12-01T22:21:22 http://permalink.gmane.org/gmane.comp.python.ipython.user/4175 What is the usage case for this. My only hesitation to add this is that there is no promise that the engine that is currently working on a task will be the engine that actually completes the task. Brian Brian Granger 2008-12-01T22:16:21 http://permalink.gmane.org/gmane.comp.python.ipython.user/4174 Yes, could you file a bug report here: https://bugs.launchpad.net/ipython Thanks! Brian On Wed, Oct 22, 2008 at 11:38 AM, Lawrence Johnston <unussum< at >gmail.com> wrote: Brian Granger 2008-12-01T22:14:46 http://permalink.gmane.org/gmane.comp.python.ipython.user/4172 Can you file a bug report on our launchpad site: https://bugs.launchpad.net/ipython Thanks, Brian On Tue, Nov 25, 2008 at 10:25 PM, Elias Bröms <elias< at >eliasit.se> wrote: Brian Granger 2008-12-01T22:12:03 http://permalink.gmane.org/gmane.comp.python.ipython.user/4171 Yep, I think that is what is going on. Is there a chance that the signal is anything other than EINTR? I ask as that will help us track this down. There are a couple of possibilities: 1. Something deep in the internals of Python itself. 2. Something deep in Twisted 3. It wouldn't be in IPython as we (as far as I know) are not sending any signals. 4. Deep somewhere in user code that they are not aware of. My best guesses are Twisted or in user code. I will look at Twisted to see if it sends signals anywhere. Is it also possible that the kernel itself sends the signal? Thanks Brian Brian Granger 2008-12-01T22:09:11 http://permalink.gmane.org/gmane.comp.python.ipython.user/4170 Yes, IPython will handle this beautifully. Here is an example, with everything done interactively: pcp025387pcs:~ bgranger$ ipython Python 2.5.1 (r251:54863, Apr 15 2008, 22:57:26) Type "copyright", "credits" or "license" for more information. IPython 0.9.1 -- An enhanced Interactive Python. ? -> Introduction and overview of IPython's features. %quickref -> Quick reference. help -> Python's own help system. object? -> Details about 'object'. ?object also works, ?? prints more. In [1]: files = ['file%i.txt' % i for i in range(10)] In [2]: files Out[2]: ['file0.txt', 'file1.txt', 'file2.txt', 'file3.txt', 'file4.txt', 'file5.txt', 'file6.txt', 'file7.txt', 'file8.txt', 'file9.txt'] In [3]: from IPython.kernel import client # this won't load balance, but has low overhead In [4]: mec = client.MultiEngineClient() In [5]: def doit(file): ...: pass ...: In [6]: mec.map(doit,files) Out[6]: [None, None, None, None, None, None, None, None, None, None] # this does load balanc Brian Granger 2008-12-01T19:00:25 http://permalink.gmane.org/gmane.comp.python.ipython.user/4169 Hi, I do have some trivial parallel code (I need to apply some analysis algorithm to a bunch of data files), and access to either an SMP linux machine with loads of CPUs or to a linux cluster. In the past, I have launched my python scripts with a different command line to get them to run concurrently, but I was thinking whether there's some pythonic way of dealing with this that is simple. I seem to remember that ipython would be a nice tool for this, but I'm not sure of it, or indeed, any other alternatives. My code is simple: for file in list_of_files: #I want spawn this next line ProcessFile ( file ) Thanks for your help, Jose Jose Gómez-Dans 2008-12-01T17:50:29 http://permalink.gmane.org/gmane.comp.python.ipython.user/4166 _______________________________________________ IPython-user mailing list IPython-user< at >scipy.org http://lists.ipython.scipy.org/mailman/listinfo/ipython-user unussum< at >gmail.com 2008-12-01T05:26:36 http://permalink.gmane.org/gmane.comp.python.ipython.user/4165 On Sun, Nov 30, 2008 at 10:13 PM, Richard Riley <rileyrgdev< at >googlemail.com> wrote: Glad that someone asked! - One reason is integration with other tools (embedding). The console takes "too much" control, and doesn't always play well with gui event loops. The "threading" versions of ipython shells (to be used with qt, gtk et al) are a constant source of woes. My particular use case for this is "ILeo": http://webpages.charter.net/edreamleo/IPythonBridge.html - Leo is being ported to Qt, and I want IPython to be first class citizen there. Sometimes the console ipython can be comfortably integrated with gui event loop using PyOS_InputHook, but esp. with Qt & windows, it was too slow. Tk is a well-behaved exception it this respect. - We can offer better feature set with gui shells. For example, the qt shell allows seamless multiline editing using the qscintilla feature set, while multiline editing with console ipython is not that stellar (unless you switch to an external editor using %edit, losing tab comple Ville M. Vainio 2008-11-30T20:49:58 http://permalink.gmane.org/gmane.comp.python.ipython.user/4164 You can now actually use the qt ui. It's at "technology demo" right now, but basic stuff like tab completion now works. The colors suck (typical ipython + white background issue), KeyboardInterrupt doesn't work (apart from doing it in the console window) and it probably has lots of other problems, but nothing that hasn't been solved in other gui's already, so it probably won't take forever to polish it. Input happens in separate scintilla widget, so it requires a bit of a mindset switch, but nothing too drastic. The big win is simplicity of multiline editing, and general "statelessness" (autoindent etc). You can test it out by doing: bzr branch lp:~villemvainio/ipython/ipython-qt-ui and running qtipy.py. Screenshot: http://img84.imageshack.us/img84/5875/ipyqteo3.png Ville M. Vainio 2008-11-30T19:28:19 http://permalink.gmane.org/gmane.comp.python.ipython.user/4163 I figured as much :-) Note that %quickref should be your first stops in problems like this. Ville M. Vainio 2008-11-30T16:48:02 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.comp.python.ipython.user